On Mind Modules, Meaning, and Wittgenstein

Noam Chomsky interviewed by Emilio Rivano

Revista de lingüística Teórica y Aplicada, 2000

The issues raised in the following exchange touch briefly and non-technically on some general topics in current linguistic theory, such as external systems, design specifications, and legibility conditions. Other topics are function, meaning, and certain perspectives on Wittgenstein. The question is raised by Emilio Rivano, the reply is by Noam Chomsky.

A few suggested readings are: Language and Thought (1994, Wakefield, R.I., and London: Moyer Bell), “Language and nature” (1995, in Mind, 104, 1-61), The Minimalist Program (1995, Cambridge, Mass.: The MIT Press), Nuestro Conocimiento del Lenguaje Humano (1998, Santiago de Chile: Ediciones Universidad de Concepción & Bravo y Allende Editores. English-Spanish Bilingual Edition), all by Noam Chomsky.

* * *

A. Emilio Rivano: There is a topic that came back to me the other day, as I was lecturing on Toulmin and Wittgenstein. It struck me also that I had mentioned the point to you in the form of a written question for the discussion session we had when you were here. And, if I am not mistaken, this is an issue you wanted to address then, but we ran out of time. The original question was this: “When you say that the language faculty produces representations conditioned by the design of other systems (the legibility conditions), what kind of conditioning is operating here? (You also speak of imposition, force and motivation in this context). And do we need these representations at the language faculty level, since we must have them at the other levels anyway (for legibility)? Wouldn’t they be produced twice?”

The topic now comes in the following form: I find it problematic, on the one hand, to conceive of human language as an organ, and, on the other, to have conditions imposed on it by external systems. Let’s suppose that “uninterpretable formal features are the mechanism that implements the displacement property”. That, to the extent that it is confirmed or accepted under current criteria, is OK. But it strikes me as a move of a different nature to go from that to conceiving “the displacement property as being forced by legibility conditions”. For it would seem to me that the legibility conditions are not “conditions” at all, but simply functions of the displacement feature on external environments.

I don’t find it problematic to conceive of human language as an organ or system, but I find it problematic to conceive of functions of this organ as another organ or system. (I think this might remove the bothersome notion of “optimal design”)

More radically: is there any need for a conceptual system? Doesn’t semantics, rather, deal with a wild variety of functions performed by the language faculty? There is a visual system, but not a system of seeing. That is, the multiplicity of visual functions cannot be the object of a theory, but, simply, or at best, of description. Vision is the organic system. The (endless variety of) visual performances, “seeing”, are functions of the system. Likewise, language is the system and “meaning” names a wild variety of functions of this system.

Roughly, this is how the topic connects with Wittgenstein. His insights, at times openly framed as frustrations about the impossibility of being part of a theory of language, move in areas of linguistic functions. And we can’t hope for any theory on functions of the human faculties, but mere descriptions. This fine analyst found himself in a world of fragments, specific meaning functions, meaningful in particular contexts; a world where no theory is possible. As the man of genius he was, he managed to say much about it, a matter we can set aside for now.

B. Noam Chomsky: The way I’ve been looking at the matter is essentially this. I think there is by now overwhelming evidence that there is a distinctive faculty of language FL, one of the many modules of mind/brain, with its own specific properties. I’m taking a “language” L to be a state of FL. L generates an infinite set of expressions EXP, each a structured complex of representations of sound and meaning, let’s say a (PF,LF) pair, “representation” of course not having the connotations of the philosophical literature and its special usage — there need not be anything (and indeed, there apparently isn’t anything) “represented” in the sense, say, of theories of ideas.

To be usable at all, the expressions (at least, a large enough set of them, which we can take to be infinite) must be accessible to the systems of language use: at least, the sensorimotor systems and the systems of thought and conceptual organization. Conventional assumptions, with traditional roots, are that PF is accessible to the sensorimotor systems, LF to the others (I think this is almost surely inadequate or wrong, but we can take it as a starting point, as has been done implicitly, in some form, for thousands of years). That is a minimal condition of usability for FL. These “external” systems (external to FL, not to the person) have their own properties, independently of language. These properties therefore impose “minimal design specifications” for FL: it must satisfy at least these, or it won’t be usable: PF and LF must be “legible” to the external systems. The “minimalist program” explores the possibility that these are also maximal conditions, in nontrivial respects — that is, that FL is in nontrivial respects an optimal solution to the minimal design specifications.

Looked at this way, the program doesn’t seem to me to be subject to the problems you raise. The only “conditioning” is that the external systems provide minimal design specifications. It’s in principle possible (though empirically presumably not) that some organism O would have something like FL (maybe our FL) that fails its legibility conditions, so that O would not be aware that it has FL, nor would any conspecific (we might discover it by some indirect means). That doesn’t seem to me problematic.

There’s no “motivation” except in the informal intuitive sense in which “functional considerations” enter into cognitive psychology and the brain sciences, and discussion of evolution. Thus, informal description often speaks of the visual system, its nature and evolution, as “motivated” by the need to perceive important properties of the world. In their serious moments, biologists understand that this is informal exposition, not to be taken literally.

There is no redundancy. Thus, the sensiromotor system has its own properties, which may or may not have been adapted for FL in the course of human evolution (there is much debate about that); same with “thought systems”. These properties impose design specifications for FL. Nothing is stated twice.

To say that “the displacement property [is] forced by legibility conditions” is to say, less informally, that expressions generated by states of FL will not be (properly) accessible to the external systems unless they satisfy this property. The thought systems, for example, require that expressions have certain properties: that they have phrases with certain relations which can be interpretated as semantic relations. Maybe agent-patient, and maybe theme-rheme, new/old information, etc. Those of the former type seem to involve “deep structure” (first Merge); those of the latter type seem to involve “displacement” (maybe the thought systems are designed to explore “edge of constructions” to detect these). If that’s the way the thought systems work, then displacement is “forced by legibility conditions” in the sense that unless FL provides for displacement, the expressions won’t be legible — properly legible, that is; they might be interpreted as some kind of word salad. Similar considerations hold at the sound side. Thus if FL provides PF without syllable structure, it could well be that the sensorimotor systems could only give partial and distorted intepretation, as a kind of noise.

Is there “any need for a conceptual system?” That seems to me a variant of an old question about whether there is thought without language. The questions are sufficiently murky so that one cannot speak with too much conviction, but it seems to me pretty clear — from introspection, efforts to understand what people are doing, comparative animal studies — that there is thought without language, and that it has systematic properties, varying with organisms. As I understand the term “conceptual system,” it is just a loose informal designation for whatever these systems turn out to be. I don’t see the slightest reason to suppose — and in fact, very much doubt — that common sense ordinary language judgments will survive inquiry into these matters (just as “life” or “motion” or “liquid” disappear as soon as the relevant sciences get beyond the most superficial levels — here I disagree strongly with the mainstream of contemporary philosophy).

“Doesn’t semantics, rather, deal with a wild variety of functions performed by the language faculty?” I think the question has to be clarified. The term “semantics” is used for all sorts of things. Thus in “formal semantics,” it is used for certain syntactic processes, at least if we adopt the conventional usage of logic and logical philosophy. In its wide-ranging uses, the term may well have the scope you suggest. All the more reason for using the term in some more precise way: personally, I prefer the way it was developed in the tradition from Peirce and Frege through Tarski-Carnap-etc., but if formal semanticists and others prefer a different usage, that’s OK, as long as one is not misled. Using the term in the Peirce etc. tradition, it’s a factual question whether natural language has “semantics”: I think there is reason to believe that it does not, as I’ve discussed elsewhere.

You’re right that the visual system (actually, systems) have all sorts of functions. Thus it seems that for mammals, including humans, there are at least two distinct visual systems extending from the primary visual cortex, one having to do with things like object recognition (apparently a “new system”), the other with guidance of action (beyond access to consciousness, and apparently an “old system,” like that of reptiles and amphibians). When one looks more closely, there are many more properties, and the informal term “function” becomes increasingly useless. As for “seeing,” that’s another informal term, dropped by the sciences as soon as they go beyond superficiality. Thus, do frogs “see”? Do we “see” with the part of our visual system that (beyond consciousness) leads us to grasp the actual object accurately when we misinterpret it (consciously) in looking at a standard perceptual illusion? These are questions of terminology, not fact. Same when we go beyond superficiality in the study of language. The term “meaning” disappears very fast, or similar notions in other languages (which are, of course, quite different in the way they handle these semantic fields). Again, I think contemporary philosophy has been much misled in these respects, taking (academic) English to be “capturing” the real world in some fundamental way — so that we have articles with titles like “the meaning of `meaning'” — a fit topic for descriptive study of some variety of English, but unlikely to be of much use in the study of the way language is used and understood.

I’d read Wittgenstein differently. I take him to be saying that we should not be misled by ordinary usage. It is fine for normal life, but science — our effort to understand the world — goes its own ways, which didn’t seem to interest him much. Of course there won’t be any theory of language in his sense, or “theory of functions of the human faculties,” just as there won’t be any theory of “the way we live.” There is no such thing as “the study of everything” in rational inquiry, including the sciences. Physics does not study everything that is going on in the world — or practically ANYTHING that is going on in the world (that’s why scientists do experiments, after all). The world is vastly too complex to study in a serious way. But I don’t see any particular insight in this; sounds to me like truism.

In fact, I’m personally disinclined to use the word “theory” outside of rather narrow branches of the empirical sciences. When people speak of “Marxist theory” or “literary theory,” my flesh creeps; I think they are mostly deluding themselves.